Model comparison
o3 vs o3-mini
o3 is the stronger model overall, scoring 47.4 to 36.7 on the Noometry Index. o3-mini costs 1.8× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Last verified . 41 shared benchmarks.
Summary
- They share 41 benchmarks with published results for both. o3 scores higher in 8 categories and o3-mini in 1 category; 9 gaps are clear of the uncertainty.
- The widest gap is in math, where o3 leads 48.9 to 27.6.
- The biggest single-benchmark swing is Fiction.LiveBench: 88.9% for o3 and 50% for o3-mini.
- o3-mini is cheaper at $1.10 / $4.40 per million input/output tokens, against $2 / $8 for o3.
Side by side
| o3 | o3-mini | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 47.4 | 36.7 |
| Released | 2025-04-16 | 2024-12-20 |
| Weights | Proprietary | Proprietary |
| Context window | 200K | 200K |
| Max output | 100K | 100K |
| Input $ / M tokens | $2 | $1.10 |
| Output $ / M tokens | $8 | $4.40 |
| Results tracked | 63 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding o3 leads
o3: 47.3 (#57), o3-mini: 41.4 (#116)
| Benchmark | o3 | o3-mini |
|---|---|---|
| Aider Polyglot | 81.3% | 60.4% |
| GSO | 8.8% | 1.3% |
| WeirdML | 52.4% | 43.7% |
| LMArena Coding | 1408 | 1378 |
| CadEval | 74% | 54% |
| SWE-bench Verified | 62.3% | — |
| SWE-bench Verified (bash only) | 58.4% | — |
| SciCode | — | 39.8% |
| LiveBench Coding | — | 82.7% |
| ALE-Bench | 933.55 | — |
Agentic & Tool Use o3 leads
o3: 34.3 (#42), o3-mini: 29.3 (#79)
| Benchmark | o3 | o3-mini |
|---|---|---|
| Berkeley Function Calling Leaderboard | 63% | — |
| GDPval | 30.8% | — |
| Cybench | — | 22.5% |
| DeepResearch Bench | 45.2% | — |
| OSWorld | 23% | — |
| LMArena Search | 1144 | — |
| METR Time Horizons | 65.4% | — |
Reasoning o3 leads
o3: 32.1 (#66), o3-mini: 16.3 (#295)
| Benchmark | o3 | o3-mini |
|---|---|---|
| ARC-AGI-2 | 6.5% | 3% |
| SimpleBench | 53.1% | 22.8% |
| ARC-AGI-1 | 60.8% | 34.5% |
| CritPt | 1.4% | 0.3% |
| Chess Puzzles | 38% | 17% |
| LMArena Hard Prompts | 1402 | 1366 |
| Mystery Game Puzzles | 29% | 7% |
| DTBench | 84.8% | 68.8% |
| LMCA | 39.7% | 19% |
| Epoch Capabilities Index | 146.88 | 140.35 |
| ForecastBench | 62.5 | 59.6 |
| Kagi LLM Benchmark | 67.6% | — |
| EnigmaEval | 13.1% | — |
| LiveBench Reasoning | — | 89.6% |
| LiveBench Data Analysis | — | 70.6% |
| LiveBench | — | 75.9% |
Math o3 leads
o3: 48.9 (#49), o3-mini: 27.6 (#236)
| Benchmark | o3 | o3-mini |
|---|---|---|
| FrontierMath (Tiers 1-3) | 33.3% | 18.6% |
| OTIS Mock AIME 2024-2025 | 84.4% | 76.9% |
| LMArena Math | 1426 | 1396 |
| MATH Level 5 | 97.8% | 96.5% |
| FrontierMath (Feb 2025 set) | 18.7% | 12.4% |
| FrontierMath Tier 4 (v1) | 2.1% | 4.2% |
| FrontierMath Tier 4 | — | 0% |
| Omni-MATH | 71.4% | — |
| LiveBench Math | — | 77.3% |
Knowledge o3 leads
o3: 54.5 (#47), o3-mini: 38.3 (#135)
| Benchmark | o3 | o3-mini |
|---|---|---|
| GPQA Diamond | 81.8% | 77% |
| SimpleQA Verified | 49.4% | 15.3% |
| Confabulations | 14.4% | 17.9% |
| LMArena Expert | 1403 | 1364 |
| Humanity's Last Exam | 20.3% | — |
| MMLU-Pro | 85.9% | — |
| GPQA (HELM) | 75.3% | — |
Multimodal Not comparable
o3: 40.7 (#35), o3-mini: —
| Benchmark | o3 | o3-mini |
|---|---|---|
| LMArena Vision | 1214 | — |
| GeoBench | 74% | — |
| VPCT | 52% | — |
Multilingual o3 leads
o3: 51.9 (#99), o3-mini: 45.9 (#159)
| Benchmark | o3 | o3-mini |
|---|---|---|
| LMArena Non-English | 1402 | 1319 |
| LMArena Chinese | 1438 | 1379 |
| LMArena French | 1430 | 1333 |
| LMArena German | 1420 | 1303 |
| LMArena Japanese | 1404 | 1286 |
| LMArena Korean | 1371 | 1314 |
| LMArena Russian | 1407 | 1304 |
| LMArena Spanish | 1395 | 1321 |
Instruction Following o3-mini leads
o3: 73.0 (#122), o3-mini: 75.3 (#67)
| Benchmark | o3 | o3-mini |
|---|---|---|
| LMArena Instruction Following | 1368 | 1338 |
| LiveBench Instruction Following | — | 84.4% |
| IFEval | 86.9% | — |
Long Context o3 leads
o3: 53.2 (#6), o3-mini: 33.8 (#251)
| Benchmark | o3 | o3-mini |
|---|---|---|
| Fiction.LiveBench | 88.9% | 50% |
| LMArena Longer Query | 1372 | 1343 |
| CL-bench | 17.8% | — |
Writing & Preference o3 leads
o3: 63.7 (#62), o3-mini: 50.4 (#177)
| Benchmark | o3 | o3-mini |
|---|---|---|
| LMArena Text | 1410 | 1337 |
| LMArena Creative Writing | 1359 | 1286 |
| Short-Story Creative Writing | 83.9% | 61.7% |
| LMArena Multi-Turn | 1405 | 1320 |
| EQ-Bench Creative Writing | 1676 | — |
| WildBench | 86.1% | — |
| LiveBench Language | — | 50.7% |
Frequently asked questions
Is o3 better than o3-mini?
o3 is the stronger model overall, scoring 47.4 to 36.7 on the Noometry Index. o3-mini costs 1.8× less per token, which makes it the better buy when o3's lead doesn't matter for your workload.
Which is cheaper, o3 or o3-mini?
o3-mini is cheaper. It lists at $1.10 per million input tokens and $4.40 per million output tokens; o3 lists at $2 and $8.
Is o3 or o3-mini better for coding?
o3 scores higher on coding benchmarks: 47.3 versus 41.4 in the Noometry coding category.
Which has the bigger context window?
Both accept 200K tokens.
How many benchmarks do o3 and o3-mini share?
41 benchmarks have published results for both models. o3 has 63 scored results on Noometry and o3-mini has 51.